Chinese artificial intelligence models have captured the majority of global usage in just months, forcing Washington to confront a fundamental shift in technological influence as the U.S. and China compete to define AI governance worldwide. On OpenRouter, Chinese models accounted for 57%-67% of tokens used in the week of September 14, up from 6%-13% in February. On Vercel, their share rose to 55% in August from 11% in January.
President Trump and Chinese leader Xi Jinping met this week and agreed to establish a formal dialogue on AI, with both sides committing to refer to the technology as "super intelligence" or "SI." Xi stated during talks that "The right thing to do on AI ... is to 'draw on each other's strengths, not guard against each other'" and proposed that "The two sides can continue their dialogue on AI, exchange views on its risks and benefits, and jointly prevent the misuse and abuse of AI." The summit produced little progress beyond establishing a bilateral communication channel and the formal dialogue commitment.
Trump said earlier this month, "We're leading China in AI...and frankly, I want to keep it that way, because whoever wins AI wins." Yet the data tells a different story. The United States maintains structural advantages—pioneering AI labs, the most advanced chips, and deep-pocketed investors—but China's models are winning on price and accessibility. American firms including Airbnb, DoorDash and Shopify have embraced Chinese models for their lower costs and greater flexibility for customization.
The Market Advantage
China's strategy centers on open-source accessibility rather than proprietary control. Eric Olander, editor in chief of The China-Global South Project, observed that "It may not be that DeepSeek or Moonshot will make much money, but the companies surrounding them – Huawei, ZTE, the others providing support, providing technology, providing all the services around AI, may do very well." This ecosystem approach creates economic incentives throughout the supply chain, not just at the model layer.
Chucheng Feng, founding partner of advisory firm Hutong Research, framed the choice facing global customers starkly: "For most people around the world, the question will ultimately be about what kind of trade-off they want to make. Do I want a model that costs 10 times as much but is six months ahead, or one that costs one-tenth as much but is only six months behind?" Peter Walker, head of insights at OpenRouter, noted that Chinese open source models released this year "can credibly perform in advanced agentic use cases, especially in regards to coding, in a way that was just not true in late 2025." Harpreet Arora, head of agentic infrastructure at Vercel, added, "Chinese models are becoming capable enough for more tasks at a much lower cost. Once a model meets the quality bar for the job, that price difference becomes compelling."
OpenRouter's data covered companies in the U.S., Europe and what it defines as the "Global South" — 82 countries across Central and South America, Africa and Asia. This geographic spread signals that China's advantage isn't limited to developing markets; it's reshaping technology adoption globally.
Competing Visions, Competing Alliances
The ideological competition mirrors the structural one. In July, Xi launched the World Artificial Intelligence Cooperation Organization, recruiting 29 countries, including Russia, Indonesia and Pakistan. This directly challenges the Pax Silica alliance that Trump formed last year to reduce reliance on China for AI supply chains. More than two dozen countries and the European Union signed up to Pax Silica, yet China's alternative vision has proven more attractive to developing economies seeking cost-effective solutions without geopolitical strings.
Daniel Remler, a senior fellow in the technology and national security program at the Center for a New American Security, warned that Chinese AI represents "real economic and security risks for the United States," cautioning that "The ultimate concern is that the integration of Chinese AI models pulls countries into a Chinese technology sphere of influence that hardens into geopolitical alignment." Remler specifically flagged Southeast Asia, noting that the region "may see significant uptake of Chinese AI models given the close economic and cultural linkages [with China] plus growing digital infrastructure," and that "Anywhere from Lagos to São Paulo to Jakarta where entrepreneurs and governments are looking for cheap, open models, will look first to Chinese AI."
George Chen, chair of digital practice at The Asia Group consultancy, explained Beijing's perspective: "Beijing continues to believe Washington seeks to contain China's rise in AI and other emerging technologies, a perception that will shape the pace and scope of future engagement for the two countries on AI."
The Intellectual Property Question
Washington has accused China of "industrial-scale theft" of American trade secrets. Anthropic and OpenAI have released reports claiming leading Chinese AI labs trained their models on outputs from American systems, a process known as distillation. Beijing has rejected those accusations. Reva Goujon, a director focusing on China at Rhodium Group, posed the critical question: "The question is what does this US administration actually do about it?" The answer remains unclear.
Autonomous Weapons and Governance Gaps
While Trump and Xi discussed AI dialogue, the United States and Russia weakened a global UN effort to regulate lethal autonomous weapons by stripping safeguards from the language. This move undercuts Washington's ability to claim moral authority on AI governance while simultaneously losing ground on market share.
Why This Matters:
The rapid adoption of Chinese AI models represents a fundamental challenge to American technological leadership and the assumptions underlying U.S. tech policy. When cost-competitive alternatives meet quality thresholds, market forces naturally favor cheaper options—a lesson that applies across industries. The fact that two major developer platforms now show Chinese models commanding 55%-67% of usage within months signals that the competitive advantage isn't theoretical; it's operational and accelerating. For policymakers focused on maintaining American influence, this presents a governance problem: blocking Chinese models requires either government intervention that disrupts markets or subsidizing American alternatives to match Chinese pricing—both costly propositions. The geopolitical dimension compounds the challenge. Countries choosing Chinese AI don't necessarily make a conscious alignment with Beijing; they're making an economic calculation. Yet once infrastructure and ecosystems build around Chinese models, switching costs rise dramatically, effectively locking in technological and potentially political relationships. The weakening of UN safeguards on autonomous weapons further complicates America's negotiating position. Without credible international agreements on killer AI, the competition devolves into a pure arms race where the fastest and cheapest solution wins, favoring the player already winning on cost. For American firms, investors, and policymakers, the question isn't whether China's rise in AI is real—the data confirms it is. The question is whether the U.S. can compete on merit or whether it will resort to restrictions that ultimately slow innovation and cede more ground to competitors.